Mediation effect of obesity on the association between triglyceride‐glucose index and hyperuricemia in Chinese hypertension adults

1 INTRODUCTION

Hyperuricemia is caused by abnormal purine metabolism, including excessive uric acid production or insufficient renal excretion, which is one of the components of metabolic syndrome. Previous studies have shown that hyperuricemia is associated with the occurrence and development of many metabolic disorders and cardiovascular diseases. For example, hypertensive patients with hyperuricemia occurred more cardiovascular events than those without hyperuricemia.1, 2 Recently epidemiological study reported that there are 170 million patients with hyperuricemia in China,3 which greatly increases the morbidity and mortality of cardiovascular events. Therefore, it is very important to optimize the risk stratification method of hyperuricemia to identify people at high risk of cardiovascular events and other complications. Some studies have reported that IR and obesity are independent risk factors for hyperuricemia, which are significantly related to the occurrence and development of hyperuricemia.4, 5

It is recognized that the bidirectional correlation between IR and hyperuricemia.6, 7 The increase of uric acid level will lead to the impairment of endothelial function, which in turn reduces insulin sensitivity by reducing the bioavailability of nitric oxide, and eventually leads to IR.8, 9 In contrast, IR induces hyperuricemia by increasing the reabsorption of uric acid and sodium in renal tubules.10 Some epidemiological studies have also confirmed the two-side effect of IR and hyperuricemia. A longitudinal study reported the unidirectional association between the two. They found that hyperuricemia leads to IR, then partially mediates the development of hypertension.11 But another study found that the improvement of insulin sensitivity reduces the level of serum uric acid.12 In short, IR and hyperuricemia promote each other to form a vicious circle, resulting in more serious organ damage. In addition, some studies confirmed that IR is related to obesity,13 and obesity is a major risk factor for hyperuricemia,14 so the correlation between IR and hyperuricemia may be partially or completely mediated by obesity.

At present, the TyG index is considered as an alternative index to identify IR,15 and it is confirmed that the correlation with the occurrence and prognosis of many IR-related diseases.16 It has been reported that there is a positive correlation between the TyG index and the risk of hyperuricemia,17-19 and the IR was more significant in patients with hypertension complicated with hyperuricemia. However, most of these studies were conducted in relatively healthy communities. The role of the TyG index in assessing the risk of hyperuricemia in hypertensive patients and the potential mechanism of obesity in it have not been fully clarified. Therefore, the purpose of our study is to explore the relationship between TyG index and the prevalence of hyperuricemia in the hypertensive population, and further to clarify whether obesity indexes (BMI, WC, HC) play an intermediary role in it, to optimize the risk stratification method of hyperuricemia and further identify people at very high risk of cardiovascular events and other complications in a hypertensive population.

2 METHODS 2.1 Study population

This was a community-based cross-sectional study, and participants were recruited from the Xinyang county, in the middle region in China from 2004 to 2005. We used a multistage cluster sample method to select a representative sample of rural community residents aged 40–75 years. A total of 13 444 patients (5270 men and 8174 women) were incorporated into the survey, which was from 63 districts of Xinyang's seven residential communities and yielding a response rate of 84.9%. Among them, 5421 hypertensive patients were identified and thoroughly examined. Hypertension was defined as diastolic blood pressure (DBP) of ≥90 mm Hg, SBP of ≥140 mm Hg, physician diagnosis, or current medication for hypertension (as defined by WHO 1999). Of 5421 hypertensive patients, 4805 patients had complete echocardiographic data, and 254 patients were excluded because of no data about other clinical characteristics or blood biochemical indexes. Ultimately, 4551 patients (1531 men and 3020 women) with integrated clinical data remained in the present study.

2.2 Clinical characteristics

Based on the age on the residence documents, we identified those who are eligible participants, then we invited them by letter or phone to the community clinic. All the participants were interviewed and required to complete a standardized questionnaire that included general information, such as sex, age, medical history, lifestyle behaviors, and so on. Anthropometric measurement was performed by experienced research staff with uniform instruments. The height and weight of all participants were measured in the upright position with light clothing and bare feet, and the error range is not more than 0.1 cm or 0.1 kg. We measured waist and hip measurements for all participants, also while standing. The Blood pressure was measured using a standard official protocol. Systolic and diastolic blood pressures (SBP, DBP) were measured using the portable Doppler device (ES-101EX, HADECO, 8 MHz probes, Kawasaki, Japan) and standard 12-cm cuff in each arm with the patients resting for at least 5 min. The Doppler Stethoscope was placed at the humeral artery fluctuation, quickly inflate the cuff to 20–30 mm Hg above the palpated SBP, and then deflated at the rate of 2–6 mm Hg/s. The first sound heard is the SBP of the brachial artery, which continues to deflate until the sound disappears or suddenly becomes weak, the scale indicated by the mercury column is DBP. The average of three readings with the participant in the sitting position after at least 5 min of rest, recorded at least 30s apart, was obtained for analysis.

Transthoracic echocardiography was performed according to standard protocols, under the supervision of two ultrasound physicians with at least 2 years of experience, and performed by two technicians trained in echocardiography at the Institute of Cardiology, Chinese Academy of Medical Sciences. The patients respiring quietly in the left decubitus position, and the echocardiographic indicators were measured at the end of systolic and end-diastolic periods of up to three cardiac cycles, including left atrium diameter, diastolic left ventricular inner diameter (LVIDD), diastolic left ventricular posterior wall thickness (PWTd), diastolic interventricular septal thickness (IVSd), E wave deceleration time, transmitted E wave velocity and transmitted A wave velocity. Then we calculated the left ventricular mass (LVM) and left ventricular mass index (LVMI) based on the above data. LVM was calculated by using the equation: 0.8×1.04 ((IVSd + LVIDD + PWTd)3—LVIDD3) + 0.6. LVMI was calculated by dividing LVM by height2.7(LVMIh2.7).

 BMI was calculated as the ratio of the weight in kg divided by the square of the height in m. Glomerular filtration rate (GFR) was used to reflect renal function, which was calculated by the Cockcroft-Gault formula: GFR (ml/min) = [(140- age)×body weight (kg)×(0.85 female)]/72×serum creatinine (mg/dl). The diagnosis of diabetes mellitus (DM) was based on an increased fasting plasma glucose (≥7.0 mmol/L), previous physician diagnosis, or current anti-diabetic medication. The diagnosis of stroke was based on the results of strict neurological examination, computed tomography, or magnetic resonance imaging tests, which were verified from local hospital records. Coronary heart disease (CAD) was diagnosed by the results of coronary arteriography, a previous myocardial infarction, or surgery or coronary revascularization.

2.3 Biochemical parameters

Fasting blood samples were obtained from an antecubital vein of participants after overnight fasting. Serum was separated from the blood samples by centrifuged on-site. Then the serum samples were delivered to the Beijing center laboratory on the dry ice for analysis. The fasting plasma glucose (FPG), triglycerides (TG), total cholesterol (TC), high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, uric acid, and other blood biochemical indices were quantified enzymatically by an automatic analyzer (Hitachi 7060, Hitachi, Tokyo, Japan).

The TyG index was calculated as ln [fasting TG (mg/dl) ×FPG (mg/dl)/2].20 Patients were divided into four groups according to the TyG index quartile: Q1 (TyG≤8.32), Q2 (8.33≤TyG≤8.66), Q3 (8.67≤TyG≤9.07), and Q4 (≥9.08). Hyperuricemia was determined as serum uric acid ≥357μmol/L (6 mg/dl) for females and ≥417μmol/L (7 mg/dl) for males.21

2.4 Statistical analysis

Data management and statistical analysis were performed using SPSS 22.0 for Windows (SPSS Inc, Chicago, IL, USA). Data are reported as the mean ± standard deviation for continuous variables and as percentages for categorical variables. Continuous variable independent sample t-test and classified variable chi-square test were used for the differences between hyperuricemia group and non-hyperuricemia group. All participants were stratified by quartiles of TyG index, baseline differences in clinical variables between groups using analysis of variance (ANOVA) for continuous variables, and Chi-squared test for categorical variables. We used the multivariable-adjusted logistic regression model to evaluate the relationship between TyG index categories and hyperuricemia, and the values of odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Sequential models were developed to minimize the effect of confounders. Model 1 was a crude model. Model 2 was adjusted for age, sex, SBP, DBP, serum creatinine, blood urea nitrogen (BUN), GFR, history of stroke, CAD, and DM. Model 3 was adjusted for all variables of model 2 plus serum TC, HDL-C, LDL-C BMI, HC, and WC.

To further examine the impact of BMI, WC, and HC on the association between TyG index and hyperuricemia, we constructed mediation models for analysis in the whole population, hyperuricemia, and non-hyperuricemia. In the model, TyG index is predictor, BMI, HC, and WC are mediators respectively, uric acid is the outcome, and the confounders in model 3 were adjusted in the mediation analysis. The simplified mediation model is presented in Figure 1, which involved several main paths as follows. Path a: the association between TyG index and obesity indexes; Path b: the association of obesity indexes with uric acid; Path c and Path c’: the total and direct effects of TyG index on uric acid, respectively. In addition, Path ab mains the indirect effect of the TyG index on uric acid, and the sum of direct and indirect effects equals the total effect. The obesity indexes played a complete mediating role in the association between the TyG index and uric acid when the total effect and indirect effect are significant but the direct effect is not significant. The obesity indexes were the incomplete mediator when the direct effect is also significant. In addition, when the indirect effect was not significant, it indicated that the obesity index did not mediate the correlation between TyG and uric acid. The differences were considered significant if a two-tailed p value < .05.

image

Mediation effect to BMI (A) or WC (B) or HC (C) on the relationship between TyG index and uric acid in the hyperuricemia group. The parameter estimate of total effect is 0.1528(0.0730–0.2325), p<.001. Adjusted for age, sex, systolic blood pressure, diastolic blood pressure, serum creatinine, blood urea nitrogen, glomerular filtration rate, the history of stroke, coronary artery disease and diabetes mellitus, serum cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol

3 RESULTS 3.1 Clinical characteristics of patients by hyperuricemia

In the whole group of 4551 hypertension patients, there were 605 patients with hyperuricemia, the prevalence of hyperuricemia was 13.29%. The clinical characteristics of the study population by hyperuricemia are described in Table 1. The average age of all participants was 58.63±8.33 years, of which patients with hyperuricemia were 60.11±8.76 years, and it was 58.41±8.24 years in people without hyperuricemia. Males comprised 33.6% of the total participants, 51.2% of the hyperuricemia group, and 30.9% of the non-hyperuricemia group. Compared to patients without hyperuricemia, those with hyperuricemia were more likely to be older, with a higher proportion of males, had higher weight, height, BMI, WC, HC, serum TG, TC, HDL-C, LDL-C, uric acid, BUN, and creatinine, and more CAD status and stroke history (all p<.05). In addition, Echocardiographic indicators (IVSd, PWTd, LVM, and LVMI) were significantly higher in the hyperuricemia group than in the non-hyperuricemia group. Particularly worth mentioning is the TyG index is significantly higher in patients with hyperuricemia than without it (8.99±0.61, 8.70±0.59, p<.001).

TABLE 1. Clinical characteristics of participants by hyperuricemia Variables Total Hyperuricemia Non-hyperuricemia p value Age (year) 58.63±8.33 60.11±8.76 58.41±8.24 <.001 Male 1531(33.6%) 310(51.2%) 1221(30.9%) <.001 Height (cm) 157.70±7.94 160.23±8.64 157.31±7.75 <.001 Weight (Kg) 65.37±14.40 69.80±11.22 64.69±14.71 <.001 BMI (kg/m2) 26.23±5.06 27.14±3.53 26.10±5.24 <.001 WC (cm) 85.45±12.09 89.68±11.92 84.80±11.98 <.001 HC (cm) 98.25±10.95 100.72±10.80 97.87±10.93 <.001 SBP (mm Hg) 163.50±24.47 165.11±25.66 163.26±24.28 .084 DBP (mm Hg) 97.04±12.62 98.18±12.88 96.86±12.57 .017 Glucose (mmol/L) 5.57±1.69 5.57±1.20 5.56±1.75 .947 Triglyceride (mmol/L) 1.68±1.24 2.17±1.52 1.61±1.17 <.001 Cholesterol (mmol/L) 5.53±1.10 5.84±1.19 5.48±1.08 <.001 HDL-C (mmol/L) 1.55±0.34 1.48±0.34 1.56±0.34 <.001 LDL-C (mmol/L) 3.15±0.86 3.37±0.93 3.12±0.84 <.001 BUN (mmol/L) 5.47±1.81 6.48±2.53 5.31±1.61 <.001 Creatinine(umol/L) 66.25±26.00 86.86±43.48 63.09±20.37 <.001 Uric acid (umol/L) 292.78±86.77 447.57±65.42 269.05±61.56 <.001 TyG index 8.73±0.60 8.99±0.61 8.70±0.59 <.001 History of stroke 467(10.3%) 83(13.7%) 384(9.7%) .002 History of CAD 415(9.1%) 85(14%) 330(8.4%) <.001 Diabetes mellitus 374(8.2%) 50(8.3%) 324(8.2%) .937 Antidiabetics 52(1.1%) 9(1.3%) 43(1.1%) .408 Lipid-lowering agent  13(0.3%) 2(0.3%) 11(0.3%) .688 Echocardiographic data IVSd (mm) 1.00±0.16 1.03±0.16 1.00±0.16 <.001 PWTd (mm) 0.97±0.14 1.00±0.14 0.97±0.13 <.001 LVM (g) 158.65±44.11 172.23±49.61 156.57±42.83 <.001 LVMIh (g/m2.7) 46.43±12.53 48.26±13.40 46.14±12.36 <.001 Abbreviations: BMI, body mass index; BUN, blood urea nitrogen; CAD, coronary artery disease, TyG index, triglyceride-glucose index; DBP, diastolic blood pressure; HC, hip circumference; HDL-C, high-density lipoprotein cholesterol; IVSd, end-diastolic interventricular septal thickness; LDL, low-density lipoprotein cholesterol; LVMIh, left ventricular mass index divided by height2.7.; PWTd, end-diastolic posterior wall thickness; SBP, systolic blood pressure; WC, waist circumference. 3.2 Clinical characteristics of patients by TyG index

Based on the above results, the TyG index was grouped by quartile, and we presented the baseline characteristics of the patients in Table 2 according to the TyG index categorical. In all participants, the prevalence of hyperuricemia with the lowest, second, third, and highest quartile of TyG index were 6.0, 10.4, 15.4, 21.4%, respectively. It means an increasing trend in the prevalence of hyperuricemia with the higher TyG index (p<.001). Compared with patients with the lowest quartiles of TyG index, those with higher quartile of TyG index had higher weight, BMI, WC, HC, SBP, DBP, heart rate, and more CAD and DM status (all p<.001). In the biochemical index results, the group with a higher TyG index had higher serum FPG, TG, TC, HDL-C, uric acid, and lower BUN, LDL-C (all p<.001). There were no significant differences in age, height, serum creatinine, or the morbidity of stroke among those groups.

TABLE 2. Clinical characteristics of participants by TyG index Variables Q1 (≤8.32) Q2 (8.33-8.66) Q3 (8.67-9.07) Q4 (≥9.08) p value N 1140 1136 1137 1138 – Age (year) 58.52±8.78 58.76±8.52 58.90±8.08 58.35±7.93 .406 Male 474(41.6%) 375(33.0%) 342(30.1%) 340(29.9%) <.001 Height (cm) 158.06±7.86 157.50±8.06 157.30±7.88 157.94±7.93 .074 Weight (Kg) 62.38±20.68 64.39±11.20 66.08±11.42 68.65±11.15 <.001 BMI (kg/m2) 24.91±7.42 25.90±3.17 26.66±3.93 27.47±3.82 <.001 WC (cm) 80.70±13.35 84.32±11.24 86.89±10.48 89.92±12.09 <.001 HC (cm) 94.12±12.89 97.84±10.18 99.71±9.26 101.32±9.78 <.001 SBP (mm Hg) 160.72±24.16 164.19±24.72 164.18±24.52 164.92±24.30 <.001 DBP (mm Hg) 95.77±12.31 97.46±13.04 97.00±12.43 97.92±12.59 <.001 Heart rate 70.74±11.28 72.06±11.70 73.97±12.65 74.51±12.99 <.001 Glucose(mmol/L) 4.90±0.67 5.20±0.72 5.46±1.00 6.71±2.75 <.001 Triglyceride (mmol/L) 0.83±0.18 1.20±0.19 1.67±0.31 3.03±1.78 <.001 Cholesterol (mmol/L) 5.05±0.90 5.36±1.00 5.70±1.05 6.02±1.18 <.001 HDL-C (mmol/L) 1.68±0.35 1.60±0.34 1.52±0.32 1.40±0.29 <.001 LDL-C (mmol/L) 2.80±0.71 3.09±0.80 3.36±0.87 3.36±0.92 <.001 Creatinine(umol/L) 66.05±26.27 66.67±22.48 66.65±30.68 65.63±23.84 .736 BUN (mmol/L) 5.63±1.83 5.54±1.77 5.40±1.84 5.30±1.76 <.001 Uric acid (umol/L) 267.72±75.74 286.00±81.65 300.58±87.62 316.85±93.48 <.001 History of stroke 102(8.9%) 110(9.7%) 132(11.6%) 123(10.8%) .160 History of CAD 78(6.8%) 93(8.2%) 95(8.4%) 149(13.0%) <.001 Diabetes mellitus 5(0.4%) 17(1.5%) 64(5.6%) 288(25.3%) <.001 Hyperuricemia 68(6.0%) 118(10.4%) 175(15.4%) 244(21.4%) <.001 Abbreviations: BMI, body mass index; BUN, blood urea nitrogen; CAD, coronary artery disease; DBP, diastolic blood pressure; HC, hip circumference; SBP, systolic blood pressure; HDL-C, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; TyG index, triglyceride-glucose index.; WC, waist circumference. 3.3 Association of TyG index with hyperuricemia

To explore the relationship between the TyG index and hyperuricemia, we built logistics regression models, and the results were presented in Table 3. Our study revealed a positive correlation between the TyG index and hyperuricemia. As shown in Table 3, participants with the second, third, and highest quartile of TyG index were 1.83, 2.87, and 4.30 times more likely to have hyperuricemia compared with the lowest quartile in the crude model (p<.001). After controlling for age, sex, SBP, DBP, serum creatinine, BUN, GFR, the history of stroke, CAD, and DM in model 2, the OR (95% CI) for hyperuricemia by TyG index groups was changed to 1.91(1.36–2.68), 3.55(2.57–4.90) and 6.29(4.56–8.68), respectively. The relationship between TyG index and hyperuricemia remained significant after further adjustment for serum TC, HDL-C, LDL-C, BMI, WC, and HC in model 3, and the OR (95% CI) of second, third, and highest quartile of TyG index compared with the lowest quartile were 1.62(1.15–2.28), 2.48(1.77–3.49) and 3.45(2.35–5.08), respectively. In particular, the association between the TyG index and hyperuricemia attenuated after being adjusted for several biochemical markers and BMI but remained significant.

TABLE 3. Odds ratio (95% CI) of hyperuricemia by triglyceride-glucose index Q1 (≤8.32) Q2 (8.33-8.66) Q3 (8.67-9.07) Q4 (≥9.08) Hyperuricemia 68(6.0%) 118(10.4%) 175(15.4%) 244(21.4%) OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value OR (95%CI) p-value Model1

Reference

1.83(1.34-2.49) <.001 2.87(2.14-3.85) <.001 4.30(3.24-5.71) <.001 Model2

Reference

1.91(1.36-2.68) <.001 3.55(2.57-4.90) <.001 6.29(4.56-8.68) <.001 Model3

Reference

1.62(1.15-2.28) .006 2.48(1.77-3.49) <.001 3.45(2.35-5.08) <.001 Abbreviations: 95% CI, 95% confidence interval; OR, odds ratio. Model 1: unadjusted. Model 2: adjusted for age, sex, systolic blood pressure, diastolic blood pressure, serum creatinine, blood urea nitrogen, glomerular filtration rate, the history of stroke, coronary artery disease and diabetes mellitus. Model 3: adjusted for model 2 plus serum cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, body mass index, waist and hip circumference. 3.4 Association of TyG index with obesity indexes and serum uric acid

Table 4 shows the associations of the TyG index with obesity indexes (BMI, WC, and HC) and serum uric acid. The β coefficients (95% CI) of second, third, highest quartile of TyG index with BMI were 0.09(0.03–0.15), 0.15(0.09–0.21), and 0.22(0.15–0.28), respectively (p < .001) after adjusting for age, sex, SBP, DBP, serum creatinine, BUN, GFR, the history of stroke, CAD and DM. Similarly. TyG index was significantly positively correlated with WC and HC. In addition, the relationship between the TyG index and serum uric acid level is also statistically significant. Compared to the lowest quartile of the TyG index, the β coefficients (95% CI) of the second, third, highest quartile of TyG index with uric acid were increased gradually, which are 0.27, 0.48, 0.74, respectively (p < .001).

TABLE 4. The association of triglyceride-glucose index with hyperuricemia and obesity indexes TyG index Q1 β(95%CI);p-value Q2 β(95%CI);p-value Q3 β(95%CI);p-value Q4 β(95%CI);p-value BMI (kg/m2) Reference 0.09(0.03-0.15); < .001 0.15(0.09-0.21); < .001 0.22(0.15-0.28); < .001 WC (cm) Reference 0.26(0.18-0.33); < .001 0.43(0.36-0.51); < .001 0.64(0.56-0.72); < .001 HC (cm) Reference 0.28(0.20-0.35); < .001 0.41(0.33-0.48); < .001 0.51(0.43-0.59); < .001 Uric acid (umol/L) Reference 0.27(0.20-0.33); < .001 0.48(0.41-0.54); < .001 0.74(0.67-0.81); < .001 Abbreviations: 95% CI, 95% confidence interval; BMI, body mass index; HC, hip circumference.; TyG index, triglyceride-glucose index; WC, waist circumference. All adjusted for age, sex, systolic blood pressure, diastolic blood pressure, serum creatinine, blood urea nitrogen, glomerular filtration rate, the history of stroke, coronary artery disease, diabetes mellitus. 3.5 The mediating role of obesity indexes

To further examine the impact of obesity indexes on the association between the TyG index and hyperuricemia, we constructed a mediation model for analysis in the whole population, hyperuricemia, and non-hyperuricemia. The results of mediation analysis were displayed in Table 5 and the simplified mediation model was presented in Figure 1 and Figures S1 and S2. We found that all of our interested obesity indexes had a mediation impact (to various extents) on the link between the TyG index and uric acid. In the total population, BMI, WC, and HC all partially mediated the correlation between TyG index and uric acid, the total effect TyG index on uric acid was 0.2446 (0.2057–0.2339), and the mediation proportion of 8.9%, 1

留言 (0)

沒有登入
gif